Considerations in the estimation of glottal volume velocity waveforms

Although numerous papers have been presented or published relating to the estimation of the glottal volume velocity waveform from speech, there is very little detailed information relating to the inherent limitations or accuracy of these procedures. When speech in an office environment is processed using reasonable quality microphones, tape recorders, analog to digital recording, and digital processing methods, the results are frequently disappointing [Holmes, J. Acoust. Soc. Am. 58, 747–749 (1975)]. The purpose of this paper is to describe several of the reasons why this disappointment occurs and to show how a combination of high quality direct recording into the computer and linear phase filtering, in conjunction with the covariance method at automatically derived anchor points, can result in theoretically expected glottal volume velocity waveforms for high intensity vowel sounds where glottal closure is known to occur. Timing relations with respect to the speech waveforms are also compared with other m...